Spotify results
Big visualization
Spotify features
All correlations
The participants view
We also left the participants with an open question to give their view on the relation between music and self esteem: ‘Do you feel like music influences your self esteem? If yes, what kind of music?’. All kinds of music were stated, while only few said that they did not think there was a causal relation. Above are the genres that were said to have a positive influence visualized, with only the genres stated that had more than two mentions. There were many more kinds of music named, including Instrumentel, Indie-pop, K-pop, Rock, Classic, Energetic, Hyped, Familiar (as in music you’re familiar with), Electronic and tunes that have a relatable context. Kinds of music that did not stimulate, but rather violate your self-esteem, were sad songs, songs about loneliness and tunes by Bon Jovi: ‘ABBA of Panic at the disco op vol volume knalt t echt even omhoog, terwijl liedjes als living on a prayer voor mij persoonlijk echt mn dag kunnen verpesten’.
Column
Hypotheses
Main hypothesis
There is a positive relationship between self-esteem and uplifting music.
Sub-hypotheses
We formulated our hypotheses based on the Spotify API features and defined uplifting music as consisting of a positive valence, high energy, high danceability, and a major mode.
As such, our sub-hypotheses are that there is a positive relationship between self-esteem and:
Danceability (\(r = .049\), \(p = .360\))
Energy (\(r = -.020\), \(p = .443\))
Mode (\(r = .016\), \(p = .453\))
Valence (\(r = .226\), \(p = .047\))
The statistical tests were done using the IBM SPSS 25 software. A one-tailed Pearson’s correlational test was conducted for the sub-hypotheses. The alpha level used was 0.05.
We also decided to explore the relationship between self-esteem and some other variables for which we did not have any prior predictions. We used two-tailed Pearson’s correlational test for these variables. The correlations between self-esteem and the other features were all non-significant, \(p > .136\). The correlation between self-esteem and general musical sophistication was also non-significant, \(p = .128\).
Spotify features
Dance – Danceability describes how suitable a track is for dancing based on a combination of musical elements including tempo, rhythm stability, beat strength, and overall regularity. A value of 0.0 is least danceable and 1.0 is most danceable.
Energy – Energy is a measure from 0.0 to 1.0 and represents a perceptual measure of intensity and activity. Typically, energetic tracks feel fast, loud, and noisy. For example, death metal has high energy, while a Bach prelude scores low on the scale. Perceptual features contributing to this attribute include dynamic range, perceived loudness, timbre, onset rate, and general entropy.
Mode – Mode indicates the modality (major or minor) of a track, the type of scale from which its melodic content is derived. Major is represented by 1 and minor is 0.
Valence – A measure from 0.0 to 1.0 describing the musical positiveness conveyed by a track. Tracks with high valence sound more positive (e.g. happy, cheerful, euphoric), while tracks with low valence sound more negative (e.g. sad, depressed, angry).
You can find out more about the different audio features on Spotify by going to this link.
Background
Column
Theory
In this modern world, music has become increasingly accessible and individualised (Skånland, 2013), through inventions such as the MP3 Player and applications like Spotify. Individuals are now able to easily tune in to their desired song or create their own playlists of songs. In comparison to the past where people heavily relied on shared music outlets like the radio. Naturally, one could request for their desired music by calling the radio hotline, but they would have to wait for their queued song, and they would need to have access to a radio. This brings us to the question of how frequent exposure to personalised music could impact individuals.
Based on previous research, we found that music can improve people’s self-esteem. One example is the study by Sharma and Jagdev (2011). The research consisted of 30 students with high academic stress and low self-esteem. They were split into two groups whereby one was selected for music therapy and the other was the control group. The participants’ stress level and self-esteem were measured using the Scale of Academic Stress and the Self Esteem Inventory. The music therapy group was instructed to listen to a 30-minute flute recording of raga, a melodic mode used in Indian classical music, daily for 15 days. The results showed that the music therapy group had higher self-esteem than the control group, which was statistically significant. This supports the idea that music therapy improves self-esteem.
Another research found that people who suffer from mental health issues, such as depression, tend to have low self-esteem and that music therapy was effective in improving their self-esteem (Hanser, & Thompson, 1994). All the research we have found with regards to music and self-esteem have been conducted in controlled environments, whereby the researchers determined the songs the participants listened to. We would like to see if the relationship between music therapy and self-esteem could also be extended to everyday music, whereby people have full control over their song selection. As there is currently no research regarding this topic, our research is exploratory and could potentially form a basis for future research into everyday music and self-esteem. If there is a correlation between everyday music and self-esteem, further research could be conducted to test if everyday music is a viable, easily accessible alternative to music therapy, for those suffering from low self-esteem.
It could be that one’s self-esteem influences the nature of the music one listens to, through influencing one’s emotional state. According to the study by Heimpel, Wood, Marshall, and Brown (2002), individuals with higher self-esteem are more likely to improve their negative moods. As such, it is plausible that people with high self-esteem experience positive moods more often than people with lower self-esteem. In addition, it was found that people tend to (consciously or unconsciously) choose music that matches their current state mood, as mentioned in the article review, and that experiencing positive emotions led to a preference for music with positive valence (Schubert, 2007). As a result, those with high self-esteem may listen to music with positive undertones more often than those with lower self-esteem.
Methods
Participants
The participants were 56 Spotify-users who filled in the “Does music define your self?” questionnaire. An additional 15 people filled in the questionnaire but were excluded from the final sample because of not submitting a link to their Spotify “Your Top Songs 2019” playlist. The sample was composed of 29 females and 27 males with a mean age of 21 years, ranging between 15 and 31. Most participants are currently in university. The highest educational level the participants finished varied.
Materials
The two questionnaires we used are discussed below:
The Rosenberg Self-Esteem Scale (Rosenberg, 1965) is a reliable measurement of one’s self-esteem (α = 0.88) (Robins, Hendin, & Trzesniewski, 2001). The questionnaire consists of 10 items rated on a four-point Likert scale from one to four. The ratings were: strongly disagree, disagree, agree and strongly agree. The lowest possible score is four and the highest is forty. The higher the score, the higher one’s level of self-esteem. An example of a test item is “I feel I do not have much to be proud of”.
The Goldsmith Musical Sophistication Index v1.0 (Müllensiefen, Gingras, Stewart, & Musil, 2013) consists of six dimensions. For our research, we opted to use only the General Musical Sophistication subscale which is a reliable (α = 0.93) measurement of one’s musical abilities and achievements. The subscale questionnaire consists of 18 items, rated on a seven-point Likert scale from one to seven. The ratings were: completely disagree, strongly disagree, disagree, neither agree nor disagree, agree, strongly agree, and completely agree. The lowest possible score is 18 and the highest is 126. The higher the score, the higher one’s level of musical sophistication. An example of a test item is “I enjoy writing about music, for example on blogs and forums.”.
Procedures
The online survey “Does music define your self?” was shared via social media. Those who filled in the survey received instructions and answered a series of questions shown on their screens. First, the participants had to paste the link to their “Your Top Songs 2019” Spotify playlist. They then answered social demographic questions regarding their gender, age, occupational status and highest achieved educational level. Following that, they answered the Gold MSI test and finally the Rosenberg self-esteem questionnaire. Finally, they could state if they thought music influences their self-esteem and indicate whether they understood the questions. None of the participants did not understand more than two questions. This is the link to our survey. Based on the Spotify API features, we used R studio to calculate the average value of each feature for each person’s playlist. The values of these features were then correlated with the self-esteem scores.
Motivation
Our group started out with an interest in everyday music as a mood booster. Through some personal reflections on our experience with music, we found a common, testable construct that has been ever-present in our interaction with music and mood - self-esteem. Self-esteem is an individual’s evaluation of their worth (Ziller, Hagey, Smith, & Long, 1969). With this in mind, we formulated our research. These are snippets of our personal motivation to conduct this research:
Annelotte
Whenever I am feeling low or down, I put on a Beyonce song and get totally lost in the music. Then it doesn’t matter anymore where I am or what I was doing, I just focus on the beats and lyrics. This always makes me forget all about why I was feeling low and gets me in the mood where I can take on the world again. That is what music is to me: helping me in my low times, but also giving a helping hand to be able to celebrate the good times even better. This made me wonder if more people listen to music to boost their confidence like I do.
Lyana
Music and self-esteem is an extremely personal topic, as I too have suffered from extremely low self-esteem. During that time, music was the only thing that truly helped me. If with this research, we could somehow contribute to an accessible solution to help those suffering from mental health and self-esteem issues, then hopefully we could help someone who is going through a painful and dark situation.
Lydia
Music has been a constant source of support in my life. Through the loneliness and struggles I’ve faced, music gave me the strength to push on. It invigorates me when I’m feeling low and allows me to wallow in my sorrow if I wish to. Music gives me comfort and confidence. For instance, when I’m walking in public without music, I feel awkward and anxious but when listening to music, I’m content and walk with confidence. As such, I was curious as to whether there’s an interactive relationship between one’s self-esteem and choice of music.
Pauline
Since I study Medicine, I am being compared to the rest of the students constantly, which results in me having low self-esteem at those times. My ultimate remedy to these times is listening to my own compiled ‘happy’ playlist. This always brings me into a happier mood. I listen to music every day, for as long as I can to keep me in my best mood. I can conclude that it made me a better person.
Pepijn
I find it fascinating that when I’m happy I tend to listen to a happy house beats that reminds me of summer, that when I’m down I put on an Ed Sheeran song about a prostitute that’s addicted to crack, and that Eminem is the one that’s fits the place when I feel like I can fight anyone. Music, at least for me, seems to reflect a state of mind, but what are the effects of reflecting this state of mind? Would it reinforce this state of mind? Would it reverse it? I find these emotional effects of music listening particularly interesting, and in the case of self-esteem these effects can be of good use, when used tactically.
Column
Our Sample
Discussion
Column
Conclusion
It has been found in multiple scientific articles that music therapy improves self-esteem. These experiments were all conducted in a controlled environment where participants could not choose what they listened to. Therefore, this exploratory research was conducted to see whether personalised everyday music has the same relationship with self-esteem as music therapy does. We hypothesized that there would be a positive correlation between self-esteem and the following Spotify features: valence, energy, danceability and mode. To test these hypotheses, a one-tailed correlation test was used. Based on the results, we can conclude that valence does indeed have positive correlation with self-esteem, as it has a significant relationship, \(p = 0.047\). Although, this is only a small to medium sized correlation, \(r = 0.226\). Contrary to our sub-hypotheses, there is no significant correlation between self-esteem and energy, danceability and mode, \(r < 0.016\), \(p = 0.443\), \(0.360\), \(0.453\). The same can be said for the other Spotify features, \(p > 0.136\), and the general music sophistication, \(p = 0.128\), where two-tailed tests were conducted instead.
As uplifting music was defined by having positive valence, high energy, high danceability, and a major mode, no clear conclusion can be made regarding the correlation between uplifting music and self-esteem. It may be that there is no positive relation between the two, contrary to our initial hypothesis. It could also be possible that the assumptions we made for uplifting music was incorrect or incomplete, as other audio features may define uplifting music better. Lastly, other features of the music, such as the message the songs are trying to convey and the use of lyrics, were not analysed in this research which may also influence self-esteem.
Number of participants
56Column
Limitations
There are a few limitations to our research:
Correlation is not causation
As much as we would have liked to conduct an experimental research, it is quite a challenge with our limited resources (such as time and a lack of compensation for the participants). Ideally, we would have liked to conduct an experiment using the daily sampling method, whereby participants would be asked to complete a survey every day regarding their mood, self-esteem and music listened to. However, that would require a lot of commitment from our participants, even if we had the appropriate incentives. As a result, we can only make correlational conclusions.
Retrospective aspect
The playlists are based on the songs someone listened to in 2019, while our research is being conducted in the present time of early 2020. It is possible that one’s self-esteem has changed since the time they listened to those songs, and thus, potentially decreased the accuracy of our findings between self-esteem and everyday music. However, the stability of one’s self-esteem gradually increases during one’s adolescence and early adulthood (Trzesniewski, Donnellan, & Robins, 2003), which is the main age group of our participants. As such, the impact of the instability of one’s self-esteem on our findings should be rather minimal.
Participant’s state
Another factor that could have resulted in inaccurate results is the state of the participants while they were completing the survey. A personal life event may have affected the person’s perception of their self-esteem. For instance, receiving bad grades could have lowered the participant’s belief in their abilities and may have temporarily lowered their self-esteem. However, since we had a substantial number of participants, these differences should average out. As such, we can have some faith in the accuracy of our findings. A way to further minimise this limitation, if we had more resources, would be to conduct the survey a few times, on different days, and take the average scores for the questionnaires.
Spotify users only
Due to the requirements of the course and the ease of collating the data, we targeted people who used Spotify regularly. As such, the data collected is probably not representative of the general population, since there may be some systematic differences between Spotify users and non-Spotify users. However, we can say that it is relatively representative of people who regularly use Spotify. Furthermore, it is appropriate and reasonable to use the Spotify population for this music-based experiment, since Spotify is the next most-used music streaming service, after Apple Music (Watson, 2020).
Future Research
To further explore the relationship found between valence and everyday music, a future longitudinal research could be conducted, whereby participants could complete the questionnaire a few times for a month, and report their “On Repeat” Spotify playlist at the end of the month. The “On Repeat” playlist consists of the most frequently played songs in the last 30 days. This research would also address the limitation of the retrospective aspect of our research as the songs collected would correspond to the period when the questionnaire was answered. In addition, this research would also reduce the effect of the participant’s state, since getting the average of a few questionnaire scores would mitigate the possible influence of any major life event.
Another aspect that could not be addressed by our exploratory research is extending medicalised results to everyday music. Another research could be conducted with a therapy group to see if everyday music is comparable to music therapy, with regards to improving self-esteem. In this experiment, the music group would be instructed to listen to their own playlists while a therapy group would listen to a playlist commonly used in music therapy. Before and after the period when the participants would listen to the music - be it a week or a month - their levels of self-esteem will be measured using a questionnaire. The number of participants with low and high self-esteem should be about the same within and between the groups. By comparing the changes in self-esteem between the two groups, it could be ascertained whether the selection of songs in music therapy is important in improving self-esteem. In addition, this research could address our correlation is not causation limitation. For the music group, the music choice of the people with low and high self-esteem could be compared. Since self-esteem is determined beforehand and the music is chosen after, if a relationship is found, it could be concluded that the difference in self-esteem resulted in choosing music with a higher or lower valence.
To see if the reverse relationship between everyday music and self-esteem exists, whereby music influences self-esteem, another research could be conducted. We would suggest selecting participants based on their level of self-esteem, using a questionnaire. We would then divide the participants into two groups where one group would listen to positive valence music and another group would listen to negative valence music, for a month. The number of participants with low and high self-esteem should be equal within and between the groups. Every participant will complete the questionnaire each week. In the end, by comparing the difference in self-esteem between the two groups, it may be concluded that listening to music with a certain valence could positively or negatively impact one’s self-esteem. It should be taken into consideration that everyone has a different taste in music. Compiling two playlists with music that every participant enjoys listening to is the greatest task in this research. Therefore, perhaps it would be good to ask the participants to put their own playlist together as long as the songs meet the valence criteria.
These are just a few future research that could be conducted based on our exploratory research. With the relevance of self-esteem in the field of clinical psychology, as well as it’s societal relevance to the general population, it would be good to further explore the relations between everyday music and self-esteem.
Bibliography
Column
Clickable bibliography
Hanser, S. B., & Thompson, L. W. (1994). Effects of a music therapy strategy on depressed older adults. Journal of gerontology, 49(6), 265-269.
Heimpel, S. A., Wood, J. V., Marshall, M. A., & Brown, J. D. (2002). Do people with low self-esteem really want to feel better? Self-esteem differences in motivation to repair negative moods. Journal of personality and social psychology, 82(1), 128.
Müllensiefen, D., Gingras, B., Stewart, L., & Musil, J. J. (2013). Goldsmiths Musical Sophistication Index (Gold-MSI) v1. 0: Technical Report and Documentation Revision 0.3. London: Goldsmiths, University of London.
Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global self-esteem: Construct validation of a single-item measure and the Rosenberg Self-Esteem Scale. Personality and social psychology bulletin, 27(2), 151-161.
Rosenberg, M. (2015). Society and the adolescent self-image. Princeton, NJ: Princeton university press.
Schubert, E. (2007). The influence of emotion, locus of emotion and familiarity upon preference in music. Psychology of Music, 35(3), 499-515.
Sharma, M., & Jagdev, T. (2011). Use of music therapy for enhancing self-esteem among academically stressed adolescents. Pakistan Journal of Psychological Research, 27(1), 53.
Skånland, M. S. (2013). Everyday music listening and affect regulation: The role of MP3 players. International journal of qualitative studies on health and well-being, 8(1), 20595.
Trzesniewski, K. H., Donnellan, M. B., & Robins, R. W. (2003). Stability of self-esteem across the life span. Journal of personality and social psychology, 84(1), 205.
Watson, A. (2020, March 11). Most popular music streaming services in the United States in March 2018 and September 2019, by monthly users.
Ziller, R. C., Hagey, J., Smith, M., & Long, B. H. (1969). Self-esteem: a self-social construct. Journal of Consulting and Clinical Psychology, 33(1), 84.
Researchers info
This research was conducted by Annelotte A. C. Bonenkamp, Lydia Tian, Maifarah L. Anthonijsz, Pauline J. Brandom Bravo Bruinsma, and Pepijn van der Klei as part of the honours course “The Data Science of Everyday Music Listening” at the University of Amsterdam, conducted by Dr. John Ashley Burgoyne.
Column
Article review
The article “Everyday music listening and affect regulation: The role of MP3 players” (Skånland, 2013) was a qualitative research that aimed to investigate the influence of listening to music, on an MP3, on a person’s emotional regulation. The importance of emotional regulation was justified by its relation to one’s well being and overall quality of life. A poor regulation of emotion was linked to mental health disorders such as depression, which has linked to a decrease in the quality of one’s life and wellbeing. This study focused on music as a potential tool for emotional regulation. In this study, there were twelve participants, six males and six females, aged 18 to 44. They were all Norwegian, relatively wealthy and assumed to have a relatively high level of education. The participants were instructed to use their MP3 players regularly, listening to their own songs. Thereafter, the researcher conducted qualitative semi-structured interviews, lasting about an hour a person. From the data gathered, the researcher formed three categories which were the use of MP3, self-regulation, and coping, but focused mostly on self-regulation.
Based on this, one of the main results was that freedom of choice was important in music being an effective affect regulator. The participants chose the music according to their desired mood state. This meant either matching their current mood or changing their mood. When they wanted to change their mood, it was usually to remove disadvantageous moods or evoke a mood that was more appropriate for the situation they were in. For example, changing a negative mood into a more positive mood to prepare for a party or social event. In general, participants usually preferred maintaining their mood rather than changing it. An interesting finding was that people did not always know what mood they were in, so they would choose the music according to “what felt right”. This helped participants to realize the mood they were in. It was suggested that music could clarify, enhance and maintain one’s mood.
Overall, it was concluded that music can be used as a way for people to balance their mood or achieve their desired mood, based on the context of the situation. The main takeaway was that people did not always aim to put themselves in a positive mood. Instead, they choose music that sustains their mood which allows them to reflect upon their feelings. It is suggested by the author that this allows people to come to terms with their feelings and accept them. To add on to the author’s conclusion, it may follow that from identifying and accepting one’s feelings, a person may find it easier to change their mood or move on to the next emotion – something akin to catharsis.
The biggest problem with this article is the lack of data. To support the author’s conclusions, the participants are quoted. It would have been good to have an appendix with the statements categorized according to the three categories identified, or at least a list of relevant statements made across the twelve individuals. In this article, it is unclear how many participants experienced the phenomena identified – such as how many participants used music to change their mood. Due to the lack of an overview of the data, it made the evidence seem a bit unreliable as it could be subjective to just one person. Learning from the difficulty in analysing and presenting the data of qualitative research as well as the limited number of participants possible, we opted for quantitative research. In addition, the findings of this article contributed to the formation of our theory.